族群聚类的大配子采样策略

Macrogamete sampling strategy of ethnic group clustering

在族群进化算法中,相对于群体,针对大配子进行族群聚类不仅可以提高族群结构的合理性,而且可以大幅降低族群聚类的计算时间。提出了参数法和排队截取法两类大配子采样策略,并对六个典型测试函数进行了仿真实验。实验结果显示,在对无约束函数的优化中,自适应参数法可产生更有效的大配子群体,能够显著提高EGEA的收敛速度和解的精度。

In ethnic group evolution algorithm （EGEA）, making ethnic group clustering in macrogamete can improve the structure of ethnic group while reducing the computing time. This paper proposed two kinds of macrogamete sampling strategies, parameter based sampling strategy and ranking truncation strategy. The simulations of unconstrained numerical optimization show self-adaptation parameter sampling strategy improve the searching efficiency of EGEA greatly.